Yeah, I was invoking Pareto. In other words, amazing how an apparently arbitrary sample ("domains with 2 letters"), as opposed to a random one, reflects the population's crappiness.